DRF¶

Supported HTTP methods and descriptions¶

URL¶

http://<h2oHost>:<h2oApiPort>/DRF.json

Input parameters¶

• classification, a boolean, <i>expert</i>

  Do classification or regression. Since version 1

• validation, a Frame, <i>expert</i>

  Validation frame. Since version 1

• n_folds, a int, <i>expert</i>

  Number of folds for cross-validation (if no validation data is specified). Since version 1

• keep_cross_validation_splits, a boolean, <i>expert</i>

  Keep cross-validation dataset splits. Since version 1

• ntrees, a int, <i>critical</i>

  Number of trees. Since version 1

• max_depth, a int, <i>critical</i>

  Maximum tree depth. Since version 1

• min_rows, a int, <i>secondary</i>

  Fewest allowed observations in a leaf (in R called ‘nodesize’). Since version 1

• nbins, a int, <i>secondary</i>

  Build a histogram of this many bins, then split at the best point. Since version 1

• score_each_iteration, a boolean, <i>expert</i>

  Perform scoring after each iteration (can be slow). Since version 1

• importance, a boolean, <i>expert</i>

  Compute variable importance (true/false).. Since version 1

• balance_classes, a boolean, <i>expert</i>

  Balance training data class counts via over/under-sampling (for imbalanced data). Since version 1

• max_after_balance_size, a float, <i>expert</i>

  Maximum relative size of the training data after balancing class counts (can be less than 1.0). Since version 1

• checkpoint, a Key, <i>expert</i>

  Model checkpoint to start building a new model from. Since version 1

• overwrite_checkpoint, a boolean, <i>expert</i>

  Overwrite checkpoint. Since version 1

• mtries, a int, <i>expert</i>

  Columns to randomly select at each level, or -1 for sqrt(#cols). Since version 1

• sample_rate, a float, <i>secondary</i>

  Sample rate, from 0. to 1.0. Since version 1

• seed, a long, <i>expert</i>

  Seed for the random number generator (autogenerated). Since version 1

• do_grpsplit, a boolean, <i>expert</i>

  Check non-contiguous group splits for categorical predictors. Since version 1

• build_tree_one_node, a boolean, <i>secondary</i>

  Run on one node only; no network overhead but fewer cpus used. Suitable for small datasets.. Since version 1

Output JSON elements¶

• xval_models, a Key[]

  Cross-validation models. Since version 1, expert

• _distribution, a long[]

  Class distribution. Since version 1, expert

• _mtry, a int

  Computed number of split features. Since version 1, expert

• _seed, a long

  Autogenerated seed. Since version 1, expert

HTTP response codes¶

200 OK Success and error responses are identical.